20cm limit-up surge. Is it the turn of the robot sector to be the cross-year main theme?

Tactile sensing becomes a "tough hurdle"

On December 4th, the concept stocks of humanoid robots reignited the A-share market. Heng'erda touched the 20% daily limit, Sanhua Intelligent Control rose more than 7%, and Ampelon, Haon Automobile Electronics, Zhejiang Rongtai, and Wuzhou Spring followed suit.

Behind the sharp rise of multiple stocks is the capital market's re - pricing of the process of "humanoid robots entering the production line". The narrative of the industry is also quietly changing, gradually shifting from "who can do better somersaults" and "whose model runs more smoothly" in previous years to who can withstand the rigors of the real factory environment.

While the whole - machine manufacturers are fiercely competing in perception algorithms, motion control systems, and large - model capabilities, the long - ignored link that truly determines whether robots can replace workers in industrial settings is the electronic skin.

This is not a metaphor. Hand tactile sense means "judgment" and "decision - making" for robots, rather than simple "execution". Without tactile sense, a robot is just a "mechanical arm" with pre - set actions; only with a precise and engineering - applicable tactile system can it truly work like a human.

In recent years, the industry's attention to tactile sensors has been continuously increasing. However, until recently, this field is still considered the most "delicate", hardest - to - tackle, and hardest - to - scale shortcoming in the industrialization of humanoid robots. For this reason, Jingzhi Sense, a subsidiary of Jinghua New Materials, has attracted extra attention from the industrial chain with its high - density fabric tactile sensing gloves. Its features such as water - washability, calibration - free, modular replaceability, and the addition of tangential sensing ability give this product the potential to leap from a "laboratory showcase" to an "industrial - line component".

01

The whole - machine races ahead, while the supply chain lags behind

Tactile sensing becomes a "tough hurdle"

Although the humanoid robot track remains highly popular in the capital market, when the whole - machine manufacturers are truly moving towards mass production, the maturity of the supply chain is significantly behind.

The competition among leading enterprises has shifted from competing in demonstration actions to truly penetrating industrial scenarios. In scenarios such as consumer electronics assembly, packaging, testing, and transfer between workstations, there are still a large number of positions relying on manual labor, with repetitive rhythms, high intensity, and high turnover rates. Enterprises are urgently in need of robots that can "truly do the work".

However, the reality is that robots can easily "dance, do handstands, and climb" in the demo environment, but once they enter the real factory environment with oil stains, dust, humidity, and high - frequency collisions, their stability drops rapidly.

Yao Maoqing, a partner of Zhiyuan Robotics, said bluntly, "Somersaults reflect the technological progress of the entire industry, but in large - scale production, people are more concerned about the value that technology can bring." He revealed that currently, whole - machine manufacturers generally send R & D personnel to stay at the suppliers' production lines, teaching drawings and adjusting codes "line by line", which is not uncommon in the industry.

Behind this phenomenon lies the most real shortcoming of the industry. The engineering capabilities of upstream components are insufficient, especially the tactile sensors responsible for direct interaction with the physical world.

The tactile system is not as easy to standardize as cameras or lidars. It is directly exposed to the external environment and constantly bears complex mechanical stresses such as wear, extrusion, slippage, and torsion. For a long time, the easy - to - damage nature, cumbersome calibration, and high cost of sensors have been recognized as industry problems. Once a fingertip part is worn or a sensing element fails, often the entire dexterous hand has to be scrapped, which is not only costly but also seriously affects the production line rhythm.

The deeper industrial contradiction is that robots need to quickly switch between different object materials and rapidly adjust the grasping force, which is a multi - modal judgment ability similar to the human hand's nervous system.

Luo Jun, the CEO of Beijing Bay Area Silicon Valley Innovation Technology, believes that it will take at least another decade for the large - scale commercialization of humanoid robots because the tactile system of the hand is too complex. Without a large amount of data and high - performance sensors, robots cannot make real - time judgments, let alone replace the subconscious force control of workers at workstations.

For this reason, tactile sensors have become the most critical and weakest link in the current industry. On the one hand, they determine the upper limit of industrial implementation; on the other hand, they are the "least mature and most fragmented" part of the supply chain. The contradiction between the two makes it the most urgent area to be broken through.

The fabric tactile solution recently launched by Jingzhi Sense aims to solve this pain point. Its water - washability allows it to remain stable in the oily, humid workshop environment; calibration - free significantly shortens the debugging cycle and improves the production line rhythm; modular disassembly reduces the scrapping cost; and the tangential sensing ability enables robots to sense slippage and prevent objects from falling. This product form, which is more oriented towards mass production and engineering, is exactly what the entire industry needs.

In a sense, the turning point of electronic skin has emerged: from "can it be made" to "can it be mass - produced, can it be affordable, and can it withstand the test".

02

The technical routes are converging

The global supply chain is being reshuffled

All along, the development of tactile sensors has not been as lively as it seems on the surface of the industry, but rather presents a situation where multiple technical routes coexist.

A research report from Shanxi Securities classifies electronic skin technologies into three major categories: piezoresistive, capacitive, and piezoelectric/visual - tactile, each with its own advantages but also facing different degrees of engineering challenges.

The visual - tactile route is "flawless" in terms of technical indicators. It captures deformation through a micro - camera inside the flexible material and then obtains force information through algorithmic decoupling, and can even grasp highly fragile items such as tofu and potato chips. However, its high cost, large volume, complex process, and difficulty in large - area installation make this route more suitable as a "demonstration technology" rather than the market mainstream.

The magnetic - sensing route based on the Hall effect performs well in six - dimensional force/torque output. Companies such as Pacini are promoting this route, but it also requires complex structural design to achieve excellent accuracy, so it is more suitable for complex force - control scenarios such as assembly and grinding.

In contrast, the piezoresistive route is considered the closest to industrial reality due to its simple structure, controllable cost, easy arraying, large - area installation, and fast response.

Tesla's Optimus adopts the piezoresistive solution, and Jingzhi Sense also chooses to innovate in fabricating based on this. An industry engineer's evaluation is very straightforward: "The piezoresistive method is not the most advanced, but it is the most likely to be mass - produced."

The "engineering realism" of the piezoresistive method is becoming an industry consensus. This is not a technical compromise but an inevitable result of the industry's transition from the era of showing off skills to the engineering era.

The convergence of technical routes is just the first - layer competition. A deeper - level competition is unfolding at the supply - chain end.

European, American, and Japanese manufacturers still dominate the electronic skin market, with the top five global players accounting for 57.1% of the market share. Chinese manufacturers are just starting out. Jingzhi Sense not only established production capacity in Dongguan but also set up a R & D center in Silicon Valley and established a cooperative relationship with the American dexterous hand manufacturer TetherIA. The founder of this company was once in charge of Tesla's Optimus dexterous hand project.

This way of "systematically binding materials - sensors - dexterous hands" is an effective path for domestic manufacturers to enter the global supply chain and a new model for future competition.

As whole - machine manufacturers have intensively announced mass - production schedules, the humanoid robot industry chain is shifting from "technology competition" to "supply - chain competition": whoever can increase the yield rate, reduce the cost, and ensure the supply stability will be able to grab a larger share of the whole - machine factory supporting in the next few years. This competition model is closer to the automotive industry chain rather than the consumer electronics industry.

According to the forecast of Shanxi Securities, the global market size of humanoid robot electronic skin will reach $756 million by 2030. Considering that electronic skin accounts for a small proportion but plays a crucial role in the BOM of the whole machine, it is very likely to become the most scarce and price - sticky upstream key component.

For the domestic supply chain, the time window is opening, and whether the technology and supply capacity can keep up with the production capacity rhythm of the whole - machine manufacturers will determine the pattern of the industry in the next five years.

03

Conclusion

The node for the industrialization of humanoid robots has come much faster than expected. Whole - machine manufacturers have shifted from the "show - off" mode to the "compete for delivery" mode. Whether humanoid robots can become real industrial productivity tools depends on whether they have a reliable, cost - controllable, and mass - producible tactile system.

An electronic skin that can be washed, does not require calibration, can be replaced by area, and can sense tangential force may not be eye - catching, but what really impresses the industrial chain is often not the fancy technology but the engineering ability.

When electronic skin moves from the laboratory to the factory, the industry will truly step into the eve of commercialization.

This article is from the WeChat official account "Songguo Finance" (ID: songguocaijing1), author: Songguo Finance, published by 36Kr with authorization.